There is a growing demand for a cheap and reliable memory device for the storage of digital information for computers, video systems, multimedia etc. This device should have a data storage capacity in excess of 10.sup.11 bytes, fast access time, high transfer rate and long term stability. Today the capacity of available digital information storage means based on optical and magnetic methods is limited to about 5.multidot.10.sup.8 bytes per square inch. Two-dimensional (2-D) memory devices such as optical and magneto-optical discs, magnetic discs and magnetic tapes are well known and represent most popular memory carriers for the storage of digital information. In optical 2-D memory devices the information is usually written as local variations of thickness, reflectivity, refractive index, or absorption coefficient of the medium. Storage devices, based on optical methods have advantages over magnetic ones because of less strict requirements of the components and environment. The possibility for parallel writing of information, i.e., simultaneous recording of information over the medium's surface, is another advantage of optical memory carriers, which is especially important for mass production. These carriers are usually formed as optical discs suitable for reading (CD-ROM) or write-once-read-many (CD-WORM) modes of operation. Their description can be found, for example, in The Compact Disc Handbook by Ken C. Pohlmann.
Unfortunately, the known-in-the-art two-dimensional optical memory devices have an important intrinsic disadvantage associated with the fact that their ultimate pixel capacity is diffraction limited by a factor of 1/.lambda..sup.2 where .lambda. is the wavelength of light employed in modern lasers. A certain increase in capacity can be achieved by special measures such as a "super resolution" at a fraction of a wavelength. However, implementation of this measure is associated with the necessity for very precise and sophisticated optical, mechanical and electronic equipment as well as in a high quality medium, which obviously makes this approach expensive and less feasible.
The main efforts to create a 3-D optical memory disc have been directed at developing of CD/CD-ROM-like optical devices, where reading is based on the modulation of a reflected beam. The modulation is the result of interference on the recorded pattern (pits) for CD/CD-ROM or variation of reflectivity for CD-R devices. The great advantages of this concept and method which has been developed since the 1970s by the electronic and computer community result in simple, reliable technology for the mass production of cheap optical carriers as well as pick-ups to play back the stored audio, video, and data information. A known method of improving the capacity of 2-D optical memory carriers is the stacking of two or more discs. The commercially available carriers implementing this approach are known as DVD's and are described, for example, in Scientific American July 1996. The disadvantage of this approach is associated with multiple reflections which occured between reflective surfaces, and led to power losses during the propagation of reading and reflected beams through the layers. The interference of light beams reflected from different layers results in beam distortions due to optical aberrations. The aberrations appear when the optical path within the storage media is changed to read different planes of stored information. High quality optical adhesives are required to assemble a stack of discs so as to reduce the influence of aberrations, bubbles, separations and inclusions and to ensure that there will not be any mechanical, thermal and chemical impact on the surface of the stacked discs. Due to the above mentioned requirements the information storage capacity of commercially available multistacked discs is limited in practice to 10.sup.10 bytes. These carriers are composed of 2 discs with 2 information layers in each. The DVD optical discs are attached together at their back sides, and it is possible to achieve a maximum total storage capacity of around 2.multidot.10.sup.10 bytes.
An alternative method of optical data storage is based on three-dimensional (3-D) recording. It is obvious that 3-D recording can dramatically increase the storage capacity of the device. There are known-in-the-art 3-D recording methods, based, for example, on 3-D volume storage by virtue of local changes of the refractive index of optical media. The 3-D writing to and reading from the bulk media has been widely reported (J. H. Stricler, W. W. Webb, Optics Lett., 16, 1970, 1991; H. Ueki, Y. Kawata, S. Kawata, Applied Optics, 35, 2457, 1996; Y. Kawata, R. Yuskaitis, T. Tanaka, T. Wilson, S. Kawata, Applied Optics, 35, 2466, 1996). The operation of such device is based on using local changes of the refractive index of the optical media. These local variations of refractive index result in the birefiingence and variations of polarization of the reading beam transmitted through the media. The variations are detectable and can be interpreted as a binary code. Among the drawbacks of this approach is the very weak signal value requiring a high power laser and highly sensitive detectors. The 3-D regular structure of the information carrier acts as a birefringent material at a macro scale introducing the non informative depolarization and defocusing of the transmitted beam. Variations of the refractive index introduce the phase modulation located in the adjacent layers, diffraction and power losses. The measurement of the transmitted beam requires two optical heads (transmitting and receiving) from both sides of the caer. This solution is very complicated and expensive since it requires simultaneous alignment of the heads to a diffraction limited spot, especially while taking into account the variation of the required optical path, medium inhomogeneity, and carrier/heads movement perturbations. The data recording is possible only in a sequence fashion, bit by bit using a laser--thus, it does not allow the implementation of cheap replication methods, such as mask lithography. The solidification/polymerization process results in uncontrollable material deformation during the recording procedure. The associated stresses may cause information distortion. Thus, all the above-mentioned drawbacks will put obstacles in the way of converting this approach to be realized into the practical 3D-memory device.
There are different works and patents related to this field, e.g., J. Russell, U.S. Pat. Nos. 4,163,600; 4,219,704; 5,278,816; and M. Best et al, U.S. Pat. Nos. 5,586,107; 5,255,262;. In order to provide a stable, reliable reading from different stacked discs, layers or surfaces, several methods have been invented, including layers with different reflective spectra, read by different wavelength lasers (Frisiem A. et al, U.S. Pat. No. 5,526,338), objective lens aligrnent from one layer to another (H. Rosen et al, U.S. Pat. No. 5,202,875), spherical aberration correction achieved simultaneously for different layer depth (W. Imaino, U.S. Pat. No. 5,373,499), changeable layer's transmission, reflectivity and/or polarization (J. Russell; U.S. Pat. No. 5,465,238), guide beam and scanned beam optical head configuration together with special guide plane utilization (A. Holtslag, U.S. Pat. No. 5,408,453). In Ota's patent U.S. Pat. No. 5,559,784 several methods of non fluorescent multilayer disc manufacturing are described, including mechanical, spin coating and photolithography techniques. A structure of multilayer optical discs is suggested, and a method of data recording in the form of refractive pits within a completely transparent material is described. This approach seems more advanced than the traditional methods using a whole reflective layer with the pits recorded as a reduction of reflection. It is stated that such a structure permits building of the multilayer disc which may be played back. However the diffraction and scattering on adjacent layers as well as birefringence are present in the disc reducing the signal to noise ratio and introducing crosstalk. Besides, the concept of pickup requires two laser sources and two optical heads which are placed on both sides of the carrier. This solution is rather complicated and difficult for implementation in a commercial device.
It should be pointed out that all multilayer data storage mediums based on the interference/reflectivity physical principles have very serious drawbacks which do not allow to provide more than two layers at high information density. An increase in amount of layers is possible only with dilution of area information density. This limits the multilayer disc capacity, which is close to DVD capacity. The data recorded on a CD/DVD disc is organized as a reflectivity variation on a data layer due to the interference between the light reflected from pit and land. Therefore, the contrast ratio for signal/noise or for "1" and "0" is limited by natural reasons (light spot size/pit size). While the layer reflectivity and thus the signal value should decrease as S.about.n.sup.-2 with the number of layers n, the scattered light noise correspondingly increases as N.about.n; thus, the signal to noise ratio decreases as S/N.about.n.sup.-3 with an increase in the number of layers n. The multiple reflection from different layers leads to multiple focuses, which obviously increases the crosstalk and reduces the signal. Furthermore, it is impossible to separate the useful signal light from the parasitic reflection and scattered light. All these drawbacks limit the practical realization of multilayer data storage down to only two layers promised by the DVD inventors and manufacturers. Further, it is difficult to expect that the great advantages of CD/DVD mass production technology based on the injection molding technique can be simply converted to fit the increased capacity multilayer disk manufacturing requirements.
Another alternative physical principle widely investigated since 1980 is based on the fluorescence phenomena. In the fluorescent memory storage the data is presented as local variations of fluorescent substance properties. The substance is illuminated with a radiation at excitation (reading) wavelength, and the fluorescence signal is registered at a different wavelength. A simple spectral filter can separate the fluorescent signal at a receiver from the noise of the excitation radiation. The fluorescent medium can provide a very high contrast ratio of 10.sup.2 -10.sup.4. The reading from a fluorescent disc is insensitive to the disc tilt. Fluorescent storage of data has a number of advantages which aid in overcoming the drawbacks of reflective, refractive and polarized memory and in building reliable, simple, cheap 3-D optical storage medium and optical head, which in principle will be much simpler than the existing CD drive.
Obvious shortcomings of the fluorescent principle are the low quantum yield of fluorescence and the low coupling efficiency of pickups.
Among the methods utilizing the fluorescence principle the most remarkable is the 3-D storage method and a rewritable optical memory based on the two-photon absorption process. This approach utilizes a fluorescent medium containing photochromic molecules capable of existing in two isomeric forms. The first isomeric form A is not fluorescent; it has absorption bands for UV radiation, and is capable of being transferred into the second isomeric form B upon the simultaneous absorption of two long wavelength photons. The second form B has an absorption band corresponding to the second harmonic of reading radiation and is capable of exhibiting fluorescence upon illumination by two photons of the reading light in the infrared range. This method is described in Peter M. Rentzepis' U.S. Pat. No. 5,268,862 where, as an active medium a dedicated photochromic material, i.e., spirobenzo-piran, maintained in a 3-D polymer matrix, is utilized.
The first disadvantage associated with 3-D memory as disclosed in Rentzepis' patent lies in the fact that the service life of the spirobenzopiran medium is limited because it is not stable at room temperatures. In order to suppress side reactions and to minimize so-called fatigue, which is defined as the gradual loss of the ability to change color after repeated "write" and "erasure" cycles, it should be stored at very low temperatures (-78.degree. C.). The second disadvantage is associated with the fact that information is to be written into the medium by a two photon absorption process, where two beams should be crossed at a very small region having dimensions in the order of magnitude .lambda..sup.3 (.about.1 .mu.m.sup.3) The medium should be illuminated by two crossed beams of infrared reading radiation to induce the fluorescence. In order to fulfill this requirement each beam should be effected by picosecond or even femtosecond pulses of light so as to ensure that the intensity is sufficient for the writing and reading cycles. The required intensity is about 10.sup.10 W/cm.sup.2. The picosecond/femtosecond lasers mentioned in Rentzepis' patent are Ti:Sapphire and Nd:YAG lasers, which are expensive devices and cannot be implemented into a cheap and compact memory setup.
It should be realized as well that illumination with two crossed beams necessitates their time synchronization so as to ensure proper overlapping of two beam pulses in space and time. In practice the optical system enabling the overlapping of two picosecond light pulses within the region of 1 .mu.m.sup.3 (and over app. 10.sup.10 pixels of carrier) is not available yet and therefore practical implementation of 2-photon absorption method in a form suggested by Rentzepis so far seems questionable.
Another problem associated with 3-D memory disclosed in Rentzepis' patent is the amount of time required for writing of information bit after bit into the disc. The required time is above 10.sup.5 sec or 25 hours, assuming feasible information writing rate of 10.sup.6 bits/sec and disc capacity of 10 GBytes. This apparently makes the disclosed 2-photon absorption method very expensive even for mass production.
In conclusion, it can be outlined that despite the known-in-the-art method of 3-D recording of information and an optical memory based thereon, nevertheless, this method does not seem a feasible one, nor is it suitable one for industrial implementation.
There are several patents describing the 2-D and 3-D multiple data layer fluorescent media structure and several methods of manufacturing such media.
The first patent describing the idea of a multilayer optical memory structure was issued to Russel, U.S. Pat. No. 4,090,031, where the possibility of using a fluorescent medium for data reproducion has also been briefly mentioned. However, in contrast to our invention, in this patent the data layers have no spacing, and they should be composed from different fluorescent materials. Furthermore, the patent requires equal track and light spot widths. The patent does discuss the required information layer structure, which is essential for multilayer memory. Furthermore, it neither describes the method of such memory manufacturing, nor the principle of operation or design of proper pickup which allows for selective reproduction of the data from such a structure. It is especially important for the multilayer memory where the interlayer crosstalk and proper selectivity means are the major problems to be solved. Consequently, the patent does not disclose practical methods for building a fluorescent 3D memory.
In Miyadera's patent No. JP 05-2768 the two data layer optical recording medium is described. These recording layers contain two different color fluorescent substances having non overlapping absorption and fluorescence bands. It allows each layer to be read separately; however, it requires a complicated optical head and complicated, if at all feasible, manufacturing process. The manufacturing process is not described.
In Chikuma's U.S. Pat. No. 5,063,556 the fluorescent recording medium is suggested where the single layer information pattern is formed from the islands of fluorescent material between a transparent substrate and reflective layer. This structure is not suitable for multilayer discs because the layer is not transparent. Moreover, the method for the formation of such fluorescent islands is not described.